A practical ranking page for general LLMs, image, video, coding, audio, OCR/document, and vision models. It prioritizes public APIs and raw CSVs, then uses gated sources as cross-checks.
| Rank | Model | Organization | Score | Access | Evidence |
|---|---|---|---|---|---|
| 1 |  claude-fable-5 Style-controlled Arena rating ↑ | anthropic | 1,509 | Proprietary | 17,799 votes Source |
| 2 | claude-opus-4-6-thinking Style-controlled Arena rating ↑ | anthropic | 1,505 | Proprietary | 67,203 votes Source |
| 3 | claude-opus-4-7-thinking Style-controlled Arena rating ↑ | anthropic | 1,502 | Proprietary | 54,781 votes Source |
| 4 | claude-opus-4-6 Style-controlled Arena rating ↑ | anthropic | 1,497 | Proprietary | 70,970 votes Source |
| 5 | qwen3.8-max Style-controlled Arena rating ↑ | alibaba | 1,496 | Proprietary | 3,327 votes Source |
| 6 | claude-opus-4-7 Style-controlled Arena rating ↑ | anthropic | 1,492 | Proprietary | 55,992 votes Source |
| 7 | claude-opus-5-high Style-controlled Arena rating ↑ | anthropic | 1,492 | Proprietary | 10,704 votes Source |
| 8 | claude-opus-5-max Style-controlled Arena rating ↑ | anthropic | 1,490 | Proprietary | 5,124 votes Source |
| 9 |  muse-spark-1.1 Style-controlled Arena rating ↑ | meta | 1,490 | Proprietary | 12,086 votes Source |
| 10 | muse-spark Style-controlled Arena rating ↑ | meta | 1,488 | Proprietary | 13,486 votes Source |
| 11 |  gemini-3-pro Style-controlled Arena rating ↑ | 1,486 | Proprietary | 41,241 votes Source | |
| 12 | gemini-3.1-pro-preview Style-controlled Arena rating ↑ | 1,485 | Proprietary | 89,133 votes Source | |
| 13 | claude-opus-4-8-thinking Style-controlled Arena rating ↑ | anthropic | 1,484 | Proprietary | 35,303 votes Source |
| 14 |  gpt-5.6-sol-xhigh Style-controlled Arena rating ↑ | openai | 1,483 | Proprietary | 10,716 votes Source |
| 15 | gemini-3.6-flash Style-controlled Arena rating ↑ | 1,483 | Proprietary | 8,575 votes Source | |
| 16 | gpt-5.5-high Style-controlled Arena rating ↑ | openai | 1,482 | Proprietary | 49,978 votes Source |
| 17 | gpt-5.4-high Style-controlled Arena rating ↑ | openai | 1,477 | Proprietary | 60,264 votes Source |
| 18 | gpt-5.5 Style-controlled Arena rating ↑ | openai | 1,476 | Proprietary | 51,172 votes Source |
| 19 | gpt-5.2-chat-latest-20260210 Style-controlled Arena rating ↑ | openai | 1,476 | Proprietary | 34,036 votes Source |
| 20 | gemini-3.5-flash-high Style-controlled Arena rating ↑ | 1,475 | Proprietary | 10,472 votes Source | |
| 21 | qwen3.7-max-preview Style-controlled Arena rating ↑ | alibaba | 1,475 | Proprietary | 3,695 votes Source |
| 22 | claude-opus-4-8 Style-controlled Arena rating ↑ | anthropic | 1,475 | Proprietary | 35,877 votes Source |
| 23 | grok-4.20-beta1 Style-controlled Arena rating ↑ | xai | 1,474 | Proprietary | 26,576 votes Source |
| 24 | gemini-3.5-flash-medium Style-controlled Arena rating ↑ | 1,474 | Proprietary | 18,778 votes Source | |
| 25 | gpt-5.5-instant Style-controlled Arena rating ↑ | openai | 1,473 | Proprietary | 25,680 votes Source |
Every source below is tagged by sync method and confidence so future Codex updates can decide what to automate or keep as a reference.
Primary mixed open/closed LLM ranking. Style control reduces preference bias caused by response formatting and tone.
Strong benchmark-based cross-check for intelligence, price, and speed. Its official API requires an API key and must remain server-side.
Best automatic source for mixed open/closed image-generation rankings.
Complements text-to-image with editing models such as GPT Image, Gemini image, Seedream, and others.
Good primary source for text-to-video, including ByteDance, Kuaishou, xAI, Google, and other providers.
Secondary video source for image-to-video workflows.
Best structured source for agentic coding on real GitHub issues.
Complements SWE-bench with web-development preference rankings.
Strong source for speech recognition. Audio generation leaderboards still need a secondary curated source.
Useful for OCR and text-centric visual understanding; treat as medium-confidence until mirrored into D1 snapshots.
Chinese OCR and document-understanding companion table.
Good live source for document-style multimodal tasks.
Broad visual reasoning and multimodal model comparison.
High-quality public leaderboard, but parsing HTML is more fragile than using benchmark APIs.
Good benchmark for editing across C++, Go, Java, JavaScript, Python, and Rust, but no stable public JSON found.